[sticky] On January 2, 2008, The U.S. National Academy of Engineering announced that Rudolf Kalman would receive the 2008 Charles Stark Draper Prize for "for the development and dissemination of the optimal digital technique (known as the Kalman Filter) that is pervasively used to control a vast array of consumer, health, commercial and defense products." The $500,000 prize was presented in Washington, DC, on February 19, 2008. More information is available here. See also the NAE press release. We have also made available some photos of the event, taken by Greg Welch.
[June 29, 2008] The June offering "Application of Kalman Filtering to GPS, INS, & Navigation" by Dr. M. S. Grewal is over. The next course will be held January 19–23, 2009, Fullerton, California, USA. Sonya says their June offering of the course "brought people from Spain, Turkey, Sweden, New Zealand, Canada, and the US. A good group, quite varied in interests and backgrounds (several PhD students), which provided for lots of questions, active discussion, and great class interaction."
On November 16, 2005, Kalman was awarded Columbia Engineering School Alumni Association’s Egleston Medal for Distinguished Engineering Achievement. See online article or local copy.
On February 19, 2008, Kalman will be awared the Charles Stark Draper Prize for "the development and dissemination of the optimal digital technique (known as the Kalman Filter) that is pervasively used to control a vast array of consumer, health, commercial and defense products." More information is available here. (See also this archived copy of the press release.)
For beginners, we highly recommend reading Chapter 1 of this book by Peter S. Maybeck on the subjects of stochastic models, estimation, and control. Although the remaining chapters may appear daunting, the book is thorough and complete.
In a 1997 Innovation column of GPS World, Larry J. Levy wrote a very nice introduction to the Kalman filter titled "The Kalman Filter: Navigation's Integration Workhorse." (Also available as PDF file.) Levy provides some historical perspective, a non-mathematical explanation, and of course a mathematical explanation with examples.
Rudolph van der Merwe maintains a very nice web site on signal processing research including work on Unscented Kalman Filters. (The idea for the UKF was originally introduced by Simon Julier and Jeff Uhlmann.) Their site contains papers and a MatLab toolkit called ReBEL which contains functions and scripts for the Kalman filter, particle filters (in general), and the Unscented Kalman Filter. (MatLab is a product of The MathWorks.) Here are the slides from a June 2002 talk Rudolph gave at a CSLU weekly seminar.
Harmonic Software sells a Kalman Filter Interface Pack (KBF) for their O-Matrix product. KBF is a GUI-based environment for graphically designing, building, and analyzing Kalman filters using the Kalman filter functions available in O-Matrix.
Andre Adrian, a Senior Engineer at DFS in Germany, used the Kalman filter (and variations) to create a central tracker for the german air traffic control. He has provided a few articles here.
OpenCV includes some Kalman Filter functions, and the Reference Manual includes some introductory prose. (The prose is quite similar to our introductory paper.) The entire library can be downloaded after agreeing to their license. The Reference Manual is in the opencv-doc package.
Michael Stevens (a Senior Research Engineer at the Australian Centre for Field Robotics) has developed a nice library of C++ Bayesian Filtering Classes. The classes represent and implement a wide variety of numerical estimation algorithms for Bayesian/Kalman Filtering. The classes provide tested and consistent numerical methods and the class hierarchy explicitly represents the variety of filtering algorithms and model types.
Alex Blocker at Boston University has developed and made available some Matlab tools for Kalman filtering, smoothing, and estimation. (See his web site for notes, instructions, and a link to the tools.) The tools are licensed under LGPL 3.0. He says that the learning algorithm he uses is similar to Kevin Murphy's, but is extended to the case of a controlled system. He includes a technical note on this algorithm and its use.
The DHI Data Assimilation in Hydrological and Hydrodynamic Models site includes their DAIHM Matlab toolbox. The toolbox is designed for data assimilation and ensemble modelling. Even though the toolbox has been developped for hydrological models, they believe it is very general and can be used in other domains/applications.
Conor Dolan has made available an executable, limited source, and documentation for a program to carry out state-space modeling based on Harvey's text. It includes facilities to model multi-subject data and multi-group data. You can get a zip file from his web site (see mkfm6 program). Note that Conor says "The source is limited to parts I wrote (in fortran). Optimization of the normal theory ML function is carried out by the NPSOL routine, which I am not at liberty to disseminate."
Dr. Stephen Pollock (Department of Economics, Queen Mary, University of London) is investigating the use of signal processing and stochastic estimation in economics. For example, he is investigating the use of filters (including the Kalman filter) on short non-stationary econometric data sequences to extract such components as the trends and seasonal fluctuations.
Jan Kybic's 1998 research report “Kalman Filtering and Speech Enhancement” details his work aimed at suppression of noise in a running car environment for hands-free mobile telephony. (Thanks to Richard Kilgour at Navman for suggesting this.)
The above represents the number of accesses since March 24, 1997. Click the left ClustrMap thumbnail to view more detailed statistics. (The right ClustrMap thumbnail is a static version from July 28, 2008—the maps are archived and reset periodically.)